-
## 🚀 Feature
I would like to request the addition of support for Recurrent Neural Networks (RNNs) in the KFACOptimizer class. Currently, the KFACOptimizer class works for linear and 2D convolution …
-
Hello.
**How to make predictions via LSTM type recurrent neural networks algorithm ???**
_We don't see anywhere in the documentation how to use LSTM-like recurrent neural networks._
**How do …
-
## ざっくり言うと
- Rush et al. (2015)(#1 で既に読んだ)のモデルを拡張.
- FFNをRNN/LSTMに変えて,Attention部分にCNNを追加
- Attentionの入力にpositional encodingを追加.
- Rush et al.を超えてSOTAを出した.
- abstractive summarization
#### キーワード
…
-
-
Hello.
**How to make predictions via LSTM type recurrent neural networks algorithm ???**
_We don't see anywhere in the documentation how to use LSTM-like recurrent neural networks._
**How do …
-
### Deep Learning Simplified Repository (Proposing new issue)
:red_circle: **Project Title** : Employee Turnover
:red_circle: **Aim** : Predict Employee Turnover with scikitlearn
:red_circle: **Dat…
-
Hi, do you continue to maintain this project? Can we accomplish the sequence_layer together?
-
### Deep Learning Simplified Repository (Proposing new issue)
:red_circle: **Project Title** : Facial Skin Diseases Classification using DL
:red_circle: **Aim** : The aim of this project is to class…
-
## 一言でいうと
文における単語間の関係(ツリー構造)を、LSTMの隠れ層で表現できるようにした研究。上位ノードが入れ替わる場合は上位配下のノード情報が全て消去される必要がある。これを表現するため、配下ノードが増える(=隠れ層の占有領域が増える)につれforgetを多くしinputを減らすmasterノードを導入している。
### 論文リンク
https://arxiv.org…
-
https://yuanmingchen.github.io/190416srnn